[FEATURE]:add vLLM preset for custom models/endpoints

[Tegrask]

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

 Add vLLM as a first-class custom model/endpoint preset, like LM Studio and Ollama.                  
                                                                                                     
 Current state:                                                                                      
 - Docs at opencode.ai/docs/providers list LM Studio, Ollama, llama.cpp, Atomic Chat, but no vLLM    
 entry.                                                                                              
 - vLLM works technically via generic custom provider (@ai-sdk/openai-compatible + options.baseURL,  
  e.g. http://localhost:8000/v1), but manual config + manual model list needed.                      
 - Old issue opencode-ai/opencode#180 is closed without preset solution.                             
 - PR #43076 mentioned provider-vllm.test.ts but was closed unmerged.                                
                                                                                                     
 Request:                                                                                            
 1. vLLM preset/docs example analogous to LM Studio/Ollama                                           
 2. Include vLLM in model auto-discovery (GET /v1/models, cf. #6231)                                 
 3. Short docs section under Providers                                                               

OpenCode version

 1.18.31                                                                                             

Operating System

 Linux